Machine-Learning

Objectives

Understand the different state-of-the-art models (linear model: regression and classification, penalization, local methods, SVM, forest); Know the use cases of these models; Know how to empirically compare different models for a given task; Know how to implement the methods studied in Python or R.

Course outline

Supervised learning; Regression; Classification; Empirical risk minimization; Model evaluation; Cross validation; Functional approximation; Model complexity; Large scale optimization; Stochastic gradient descent; Regularization; RIDGE and LASSO; Support Vector Machine; Kernel trick; Ensemble methods; Aggregation and Boosting; Random forest.

Prerequisites

Regression, logistic regression, convergence of probability distributions, central limit theorem, linear algebra, tree classification, programming with R; object-oriented programming, optimization: maximization of a function, Newton’s algorithm.